US7016502B2 - Encoder and decoder - Google Patents

Encoder and decoder Download PDF

Info

Publication number
US7016502B2
US7016502B2 US10/204,298 US20429802A US7016502B2 US 7016502 B2 US7016502 B2 US 7016502B2 US 20429802 A US20429802 A US 20429802A US 7016502 B2 US7016502 B2 US 7016502B2
Authority
US
United States
Prior art keywords
encoding
accuracy information
quantization accuracy
normalization
differential value
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Lifetime, expires
Application number
US10/204,298
Other languages
English (en)
Other versions
US20030112979A1 (en
Inventor
Keisuke Toyama
Shiro Suzuki
Minoru Tsuji
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Sony Corp
Original Assignee
Sony Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Sony Corp filed Critical Sony Corp
Assigned to SONY CORPORATION reassignment SONY CORPORATION ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: SUZUKI, SHIRO, TOYAMA, KEISUKE, TSUJI, MINORU
Publication of US20030112979A1 publication Critical patent/US20030112979A1/en
Priority to US11/324,633 priority Critical patent/US7933417B2/en
Application granted granted Critical
Publication of US7016502B2 publication Critical patent/US7016502B2/en
Adjusted expiration legal-status Critical
Expired - Lifetime legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/02—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using spectral analysis, e.g. transform vocoders or subband vocoders
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B1/00—Details of transmission systems, not covered by a single one of groups H04B3/00 - H04B13/00; Details of transmission systems not characterised by the medium used for transmission
    • H04B1/66—Details of transmission systems, not covered by a single one of groups H04B3/00 - H04B13/00; Details of transmission systems not characterised by the medium used for transmission for reducing bandwidth of signals; for improving efficiency of transmission
    • H04B1/667—Details of transmission systems, not covered by a single one of groups H04B3/00 - H04B13/00; Details of transmission systems not characterised by the medium used for transmission for reducing bandwidth of signals; for improving efficiency of transmission using a division in frequency subbands
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/02—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using spectral analysis, e.g. transform vocoders or subband vocoders
    • G10L19/0204—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using spectral analysis, e.g. transform vocoders or subband vocoders using subband decomposition
    • G10L19/0208—Subband vocoders

Definitions

  • the present invention relates to encoding devices and decoding devices, and more particularly relates to an encoding device capable of improving the encoding efficiency and to a decoding device.
  • FIG. 1 shows an example of the configuration of a known encoding device.
  • a band separator 1 separates an input audio signal into a plurality of frequency bands (27 bands in this case) and outputs signals from resultant encoding units A 0 to A 26 to corresponding normalizers 2 - 1 to 2 - 27 and to a quantization accuracy determination unit 3 in each predetermined time block (frame).
  • the encoding units are simply referred to as the “encoding units A”. Similar notations are used in other cases.
  • the band separation by the band separator 1 can be performed by using a filter such as a QMF (Quadrature Mirror Filter) or a PQF (Polyphase Quadrature Filter) or by grouping spectral signals generated by spectral transformation such as MDCT in units of bands.
  • a filter such as a QMF (Quadrature Mirror Filter) or a PQF (Polyphase Quadrature Filter) or by grouping spectral signals generated by spectral transformation such as MDCT in units of bands.
  • the encoding units A can be of the same length or can have different lengths, depending on a critical bandwidth.
  • the normalizer 2 - 1 detects the signal component having the largest absolute value from the signal from the encoding unit A 0 from the band separator 1 and computes a normalization coefficient B 0 for the encoding unit A 0 using the detected value.
  • the normalizer 2 - 1 outputs the computed normalization coefficient B 0 to a multiplexer 5 , and also normalizes the signal of the encoding unit A 0 in accordance with a value corresponding to the normalization coefficient B 0 and outputs the resultant normalized data C 0 to a quantizer 4 - 1 .
  • the normalizers 2 - 2 to 2 - 27 compute normalization coefficients B 1 to B 26 based on the signals from the encoding units A 1 to A 26 and output the normalization coefficients B 1 to B 26 to the multiplexer 5 . Also, the normalizers 2 - 2 to 2 - 27 generate normalized data C 1 to C 26 and output the normalized data C 1 to C 26 to quantizers 4 - 2 to 4 - 27 .
  • the normalized data C are values within the range from ⁇ 1.0 to 1.0.
  • the quantization accuracy determination unit 3 determines quantization steps for quantizing the normalized data C 0 to C 26 in accordance with the signals of the encoding units A 0 to A 26 from the band separator 1 and outputs quantization accuracy information D 0 to D 26 , corresponding to the determined quantization steps, to the corresponding quantizers 4 - 1 to 4 - 27 .
  • the quantization accuracy determination unit 3 also outputs the quantization accuracy information D 0 to D 26 to the multiplexer 5 .
  • the quantizer 4 - 1 quantizes the normalized data C 0 from the normalizer 2 - 1 using the quantization step corresponding to the quantization accuracy information D 0 from the quantization accuracy determination unit 3 and outputs a resultant quantization coefficient F 0 to the multiplexer 5 .
  • the quantizers 4 - 2 to 4 - 27 quantize the normalized data C 1 to C 26 using the quantization steps corresponding to the quantization accuracy information D 1 to D 26 from the quantization accuracy determination unit 3 and output resultant quantization coefficients F 1 to F 26 to the multiplexer 5 .
  • the multiplexer 5 encodes the normalization coefficients B 0 to B 26 from the normalizers 2 - 1 to 2 - 27 into 6-bit codes and the quantization accuracy information D 0 to D 26 from the quantization accuracy determination unit 3 into 3-bit codes. Also, the multiplexer 5 encodes the quantization coefficients F 0 to F 26 from the quantizers 4 - 1 to 4 - 27 , multiplexes various resultant data generated by encoding, and generates encoded data. The processing of the multiplexer 5 is performed in units of time blocks (frames).
  • FIG. 2 shows an example of the configuration of a decoding device for decoding the encoded data generated by the encoding device shown in FIG. 1 .
  • a demultiplexer 21 decodes the encoded data, which is supplied from the encoding device, into the normalization coefficients B 0 to B 26 , the quantization accuracy information D 0 to D 26 , and the quantization coefficients F 0 to F 26 , and outputs the separated pieces of data to corresponding signal component composers 22 - 1 to 22 - 27 .
  • the signal component composer 22 - 1 dequantizes the quantization coefficient F 0 from the demultiplexer 21 in accordance with the quantization step corresponding to the quantization accuracy information D 0 , thus generating the normalized data C 0 . Also, the signal component composer 22 - 1 multiplies (denormalizes) the normalized data C 0 by a value corresponding to the normalization coefficient B 0 , thus decoding the signal of the encoding unit A 0 , and outputs the decoded signal to a band combiner 23 .
  • the signal component composers 22 - 2 to 22 - 27 dequantize the quantization coefficients F 1 to F 26 from the demultiplexer 21 in accordance with the quantization steps corresponding to the quantization accuracy information D 1 to D 26 , thus generating the normalized data C 1 to C 26
  • the signal component composers 22 - 2 to 22 - 27 multiply (denormalize) the normalized data C 1 to C 26 by values corresponding to the normalization coefficients B 1 to B 26 , thus decoding the signals of the encoding units A 1 to A 26 , and output the decoded signals to the band combiner 23 .
  • the band combiner 23 combines the bands of the signals of the encoding units A 0 to A 26 from the signal component composers 22 - 1 to 22 - 27 , thereby decoding the original audio signal.
  • the normalization coefficient B corresponding to one encoding unit A is uniformly encoded into, for example, 6-bit data.
  • the quantization accuracy information D corresponding to one encoding unit A is uniformly encoded into, for example, 3-bit data.
  • a first encoding device of the present invention includes band separating means for separating an input acoustic time-series signal into N bands and generating N band signals; normalization coefficient generating means for generating a predetermined normalization coefficient for each of the band signals; normalization means for normalizing the band signals on the basis of the generated normalization coefficients; quantization accuracy information generating means for generating quantization accuracy information for each of the band signals; quantization means for quantizing each of the normalized band signals on the basis of the quantization accuracy information; a plurality of normalization coefficient encoding means for encoding the N normalization coefficients, which are generated by the normalization coefficient generating means; first selection means for selecting one of the normalization coefficient encoding means on the basis of the amount of encoding in a case in which the N normalization coefficients are encoded by each of the normalization coefficient encoding means; and multiplexing means for multiplexing the N normalization coefficients, which are encoded using the normalization coefficient encoding means selected by the first selection means, and each of the
  • the band separating means may generate the N band signals by converting the input acoustic time-series signal into frequency components and separating a resultant spectral signal into N bands.
  • At least one of the normalization coefficient encoding means may include second selection means for selecting L indexes at high frequencies from among indexes of the N normalization coefficients; computation means for detecting a maximum value and a minimum value of the L selected indexes and computing the difference between the maximum value and the minimum value; encoding means for encoding each of the L normalization coefficients at high frequencies by representing, using a predetermined number of bits, a value obtained by subtracting the minimum value from each of the L selected indexes at high frequencies; and output means for outputting, when the normalization coefficient encoding means is selected, information indicating that the L indexes are specified, the predetermined number of bits, the minimum value, indexes of (N-L) normalization coefficients unselected by the selection means, and the L encoded normalization coefficients.
  • the encoding device may further include weighting means for adding a predetermined weight value to each index of the normalization coefficients so as to increase the correlation between the indexes of the normalization coefficients.
  • the encoding means may encode the indexes weighted by the weighting means.
  • At least one of the normalization coefficient encoding means may include differential value computing means for computing a differential value between an index of each of the normalization coefficients and an index of another normalization coefficient that is highly correlated with the index of each of the normalization coefficients; and encoding means for encoding the differential value, which is computed by the differential value computing means.
  • the normalization coefficient encoding means may include table maintaining means for maintaining a table in which a differential value having a high frequency of occurrence of being computed by the differential value computing means is associated with a code having a small number of bits.
  • the encoding means may encode the differential value, which is computed by the differential value computing means, by reading, from the table, a code corresponding to the differential value.
  • the table may include codes corresponding to all differential values that can be computed by the computing means.
  • the table may only include a code corresponding to a differential value having a high frequency of occurrence.
  • the encoding means may encode the differential value by outputting an index for the differential value subsequent to a predetermined escape code.
  • the encoding means may repeatedly output a second escape code, the number of repetitions being based on the size of the absolute value of the differential value, subsequent to a first escape code that depends on the sign of the differential value and may read, from the table, a code corresponding to a value obtained by subtracting a predetermined number based on the size of the absolute value of the differential value from the differential value to encode the differential value.
  • the encoding means may read, from the table, a code corresponding to a value obtained by subtracting a predetermined number that depends on the sign of the differential value from the differential value and may output the code subsequent to a first escape code, and, when the differential value other than that with a high frequency of occurrence is not within the predetermined range, the encoding means may output an index for the differential value subsequent to a second escape code to encode the differential value.
  • a normalization coefficient in a band adjacent to a band corresponding to each of the normalization coefficients may be used as another normalization coefficient highly correlated with each of the normalization coefficients.
  • the normalization coefficient encoding means may not encode differential values at frequencies higher than the particular band.
  • the normalization coefficient encoding means may make the code length of each of the normalization coefficients at frequencies higher than the particular band a predetermined value smaller than the code length of the normalization coefficient at frequencies lower than the particular band.
  • the encoding device may further include weighting means for adding a predetermined weight value to each index of the normalization coefficients so as to increase the correlation between the indexes of the differential values between the normalization coefficients.
  • the encoding means may encode the indexes weighted by the weighting means.
  • the weighting means may add a step-by-step weight that gradually increases as the band becomes higher.
  • a normalization coefficient that is temporarily adjacent to a band corresponding to each of the normalization coefficients may be used as another normalization coefficient highly correlated with each of the normalization coefficients.
  • the acoustic time-series signal may be a left signal or a right signal of a stereo audio signal.
  • the differential value computing means may compute the differential value between a normalization coefficient of the left signal and a normalization coefficient of the right signal.
  • a first encoding method of the present invention includes a band separating step of separating an input acoustic time-series signal into N bands and generating N band signals; a normalization coefficient generating step of generating a predetermined normalization coefficient for each of the band signals; a normalization step of normalizing the band signals on the basis of the generated normalization coefficients; a quantization accuracy information generating step of generating quantization accuracy information for each of the band signals; a quantization step of quantizing each of the normalized band signals on the basis of the quantization accuracy information; a plurality of normalization coefficient encoding steps of encoding the N normalization coefficients, which are generated in the normalization coefficient generating step; a selection step of selecting one of the normalization coefficient encoding steps on the basis of the amount of encoding in a case in which the N normalization coefficients are encoded in each of the normalization coefficient encoding steps; and a multiplexing step of multiplexing the N normalization coefficients, which are encoded using the normalization coefficient en
  • a first recording medium of the present invention has a program recorded therein, the program including a band separating step of separating an input acoustic time-series signal into N bands and generating N band signals; a normalization coefficient generating step of generating a predetermined normalization coefficient for each of the band signals; a normalization step of normalizing the band signals on the basis of the generated normalization coefficients; a quantization accuracy information generating step of generating quantization accuracy information for each of the band signals; a quantization step of quantizing each of the normalized band signals on the basis of the quantization accuracy information; a plurality of normalization coefficient encoding steps of encoding the N normalization coefficients, which are generated in the normalization coefficient generating step; a selection step of selecting one of the normalization coefficient encoding steps on the basis of the amount of encoding in a case in which the N normalization coefficients are encoded in each of the normalization coefficient encoding steps; and a multiplexing step of multiplexing the N normalization coefficients, which are encode
  • an input acoustic time-series signal is separated into N bands to generate N band signals.
  • a predetermined normalization coefficient for each of the band signals is generated.
  • the band signals are normalized on the basis of the generated normalization coefficients.
  • Quantization accuracy information for each of the band signals is generated.
  • Each of the normalized band signals is quantized on the basis of the quantization accuracy information.
  • the N generated normalization coefficients are encoded.
  • One of normalization coefficient encoding methods is selected on the basis of the amount of encoding in a case in which the N normalization coefficients are encoded.
  • the N normalization coefficients, which are encoded using the selected normalization coefficient encoding method, and each of the band signals are multiplexed.
  • a second encoding device of the present invention includes band separating means for separating an input acoustic time-series signal into N bands and generating N band signals; normalization coefficient generating means for generating a predetermined normalization coefficient for each of the band signals; normalization means for normalizing the band signals on the basis of the generated normalization coefficients; quantization accuracy information generating means for generating quantization accuracy information for each of the band signals; quantization means for quantizing each of the normalized band signals on the basis of the quantization accuracy information; a plurality of quantization accuracy information encoding means for encoding the N pieces of quantization accuracy information, which are generated by the quantization accuracy information generating means; first selection means for selecting one of the quantization accuracy information encoding means on the basis of the amount of encoding in a case in which the N pieces of quantization accuracy information are encoded by each of the quantization accuracy information encoding means; and multiplexing means for multiplexing the N pieces of quantization accuracy information, which are encoded using the quantization accuracy information encoding means
  • the band separating means may generate the N band signals by converting the input acoustic time-series signal into frequency components and separating a resultant spectral signal into N bands.
  • At least one of the quantization accuracy information encoding means may include second selection means for selecting L indexes at high frequencies from among indexes of the N pieces of quantization accuracy information; computation means for detecting a maximum value and a minimum value of the L selected indexes and computing the difference between the maximum value and the minimum value; encoding means for encoding each of the L normalization coefficients at high frequencies by representing, using a predetermined number of bits, a value obtained by subtracting the minimum value from each of the L selected indexes at high frequencies; and output means for outputting, when the quantization accuracy information encoding means is selected, information indicating that the L indexes are specified, the predetermined number of bits, the minimum value, indexes of (N-L) pieces of quantization accuracy information unselected by the selection means, and the L pieces of encoded quantization accuracy information.
  • the encoding device may further include weighting means for adding a predetermined weight value to each index of the pieces of quantization accuracy information so as to increase the correlation between the indexes of the pieces of quantization accuracy information.
  • the encoding means may encode the indexes weighted by the weighting means.
  • At least one of the quantization accuracy information encoding means may include differential value computing means for computing a differential value between an index of each of the pieces of quantization accuracy information and an index of another piece of quantization accuracy information that is highly correlated with the index of each of the pieces of quantization accuracy information; and encoding means for encoding the differential value, which is computed by the differential value computing means.
  • the quantization accuracy information encoding means may include table maintaining means for maintaining a table in which a differential value having a high frequency of occurrence of being computed by the differential value computing means is associated with a code having a small number of bits.
  • the encoding means may encode the differential value, which is computed by the differential value computing means, by reading, from the table, a code corresponding to the differential value.
  • the table may include codes corresponding to all differential values that can be computed by the computing means.
  • the table may only include a code corresponding to a differential value having a high frequency of occurrence.
  • the encoding means may encode the differential value by outputting an index for the differential value subsequent to a predetermined escape code.
  • the encoding means may repeatedly, output a second escape code, the number of repetitions being based on the size of the absolute value of the differential value, subsequent to a first escape code that depends on the sign of the differential value and may read, from the table, a code corresponding to a value obtained by subtracting a predetermined number based on the size of the absolute value of the differential value from the differential value to encode the differential value.
  • the encoding means may read, from the table, a code corresponding to a value obtained by subtracting a predetermined number that depends on the sign of the differential value from the differential value and outputs the code subsequent to a first escape code, and, when the differential value other than that with a high frequency of occurrence is not within the predetermined range, the encoding means may output an index for the differential value subsequent to a second escape code to encode the differential value.
  • Quantization accuracy information in a band adjacent to a band corresponding to each of the pieces of quantization accuracy information may be used as another piece of quantization accuracy information highly correlated with each of the pieces of quantization accuracy information.
  • the quantization accuracy information encoding means may not encode differential values at frequencies higher than the particular band.
  • the quantization accuracy information encoding means may make the code length of each of the pieces of quantization accuracy information at frequencies higher than the particular band a predetermined value smaller than the code length of the quantization accuracy information at frequencies lower than the particular band.
  • the encoding device may further include weighting means for adding a predetermined weight value to each index of the pieces of quantization accuracy information so as to increase the correlation between the indexes of the differential values between the pieces of quantization accuracy information.
  • the encoding means may encode the indexes weighted by the weighting means.
  • the weighting means may add a step-by-step weight that gradually increases as the band becomes higher.
  • Quantization accuracy information that is temporarily adjacent to a band corresponding to each of the pieces of quantization accuracy information may be used as another piece of quantization accuracy information highly correlated with each of the pieces of quantization accuracy information.
  • the acoustic time-series signal may be a left signal or a right signal of a stereo audio signal.
  • the differential value computing means may compute the differential value between quantization accuracy information of the left signal and quantization accuracy information of the right signal.
  • a second encoding method of the present invention includes a band separating step of separating an input acoustic time-series signal into N bands and generating N band signals; a normalization coefficient generating step of generating a predetermined normalization coefficient for each of the band signals; a normalization step of normalizing the band signals on the basis of the generated normalization coefficients; a quantization accuracy information generating step of generating quantization accuracy information for each of the band signals; a quantization step of quantizing each of the normalized band signals on the basis of the quantization accuracy information; a plurality of quantization accuracy information encoding steps of encoding the N pieces of quantization accuracy information, which are generated in the quantization accuracy information generating step; a selection step of selecting one of the quantization accuracy information encoding steps on the basis of the amount of encoding in a case in which the N pieces of quantization accuracy information are encoded in each of the quantization accuracy information encoding steps; and a multiplexing step of multiplexing the N pieces of quantization accuracy information, which
  • a second recording medium of the present invention has a program recorded therein, the program including a band separating step of separating an input acoustic time-series signal into N bands and generating N band signals; a normalization coefficient generating step of generating a predetermined normalization coefficient for each of the band signals; a normalization step of normalizing the band signals on the basis of the generated normalization coefficients; a quantization accuracy information generating step of generating quantization accuracy information for each of the band signals; a quantization step of quantizing each of the normalized band signals on the basis of the quantization accuracy information; a plurality of quantization accuracy information encoding steps of encoding the N pieces of quantization accuracy information, which are generated in the quantization accuracy information generating step; a selection step of selecting one of the quantization accuracy information encoding steps on the basis of the amount of encoding in a case in which the N pieces of quantization accuracy information are encoded in each of the quantization accuracy information encoding steps; and a multiplexing step of multiplexing the N
  • an input acoustic time-series signal is separated into N bands to generate N band signals.
  • a predetermined normalization coefficient for each of the band signals is generated.
  • the band signals are normalized on the basis of the generated normalization coefficients.
  • Quantization accuracy information for each of the band signals is generated.
  • Each of the normalized band signals is quantized on the basis of the quantization accuracy information.
  • the N pieces of generated quantization accuracy information are encoded.
  • One of quantization accuracy information encoding methods is selected on the basis of the amount of encoding in a case in which the N pieces of quantization accuracy information are encoded.
  • the N pieces of quantization accuracy information, which are encoded using the selected quantization accuracy information encoding method, and each of the band signals are multiplexed.
  • a first decoding device of the present invention includes demultiplexing means for demultiplexing the multiplexing of at least one band signal with quantization accuracy information and with an encoded normalization coefficient for each band signal; normalization coefficient decoding means for decoding the demultiplexed normalization coefficient; signal generating means for dequantizing each band signal on the basis of the quantization accuracy information and for denormalizing the dequantized signal on the basis of the normalization coefficient, whereby an acoustic time-series signal is generated; and combining means for combining bands of band signals generated by the signal generating means.
  • a first decoding method of the present invention includes a demultiplexing step of demultiplexing the multiplexing of at least one band signal with quantization accuracy information and with an encoded normalization coefficient for each band signal; a normalization coefficient decoding step of decoding the demultiplexed normalization coefficient; a signal generating step of dequantizing each band signal on the basis of the quantization accuracy information and for denormalizing the dequantized signal on the basis of the normalization coefficient, whereby an acoustic time-series signal is generated; and a combining step of combining bands of band signals generated in the signal generating step.
  • a third recording medium of the present invention has a program recorded therein, the program including a demultiplexing step of demultiplexing the multiplexing of at least one band signal with quantization accuracy information and with an encoded normalization coefficient for each band signal; a normalization coefficient decoding step of decoding the demultiplexed normalization coefficient; a signal generating step of dequantizing each band signal on the basis of the quantization accuracy information and for denormalizing the dequantized signal on the basis of the normalization coefficient, whereby an acoustic time-series signal is generated; and a combining step of combining bands of band signals generated in the signal generating step.
  • the multiplexing of at least one band signal with quantization accuracy information and with an encoded normalization coefficient for each band signal is demultiplexed.
  • the demultiplexed normalization coefficient is decoded.
  • Each band signal is dequantized on the basis of the quantization accuracy information, and the dequantized signal is denormalized on the basis of the normalization coefficient, whereby an acoustic time-series signal is generated. Bands of generated band signals are combined.
  • a second decoding device of the present invention includes demultiplexing means for demultiplexing the multiplexing of at least one band signal with a normalization coefficient and with encoded quantization accuracy information for each band signal; quantization accuracy information decoding means for decoding the demultiplexed quantization accuracy information; signal generating means for dequantizing each band signal on the basis of the quantization accuracy information and for denormalizing the dequantized signal on the basis of the normalization coefficient, whereby an acoustic time-series signal is generated; and combining means for combining bands of band signals generated by the signal generating means.
  • a second decoding method of the present invention includes a demultiplexing step of demultiplexing the multiplexing of at least one band signal with a normalization coefficient and with encoded quantization accuracy information for each band signal; a quantization accuracy information decoding step of decoding the demultiplexed quantization accuracy information; a signal generating step of dequantizing each band signal on the basis of the quantization accuracy information and for denormalizing the dequantized signal on the basis of the normalization coefficient, whereby an acoustic time-series signal is generated; and a combining step of combining bands of band signals generated in the signal generating step.
  • a fourth recording medium of the present invention has a program recorded therein, the program including a demultiplexing step of demultiplexing the multiplexing of at least one band signal with a normalization coefficient and with encoded quantization accuracy information for each band signal; a quantization accuracy information decoding step of decoding the demultiplexed quantization accuracy information; a signal generating step of dequantizing each band signal on the basis of the quantization accuracy information and for denormalizing the dequantized signal on the basis of the normalization coefficient, whereby an acoustic time-series signal is generated; and a combining step of combining bands of band signals generated in the signal generating step.
  • the multiplexing of at least one band signal with a normalization coefficient and with encoded quantization accuracy information for each band signal is demultiplexed.
  • the demultiplexed quantization accuracy information is decoded.
  • Each band signal is dequantized on the basis of the quantization accuracy information, and the dequantized signal is denormalized on the basis of the normalization coefficient, whereby an acoustic time-series signal is generated. Bands of generated band signals are combined.
  • FIG. 1 is a block diagram showing an example of the configuration of a known encoding device.
  • FIG. 2 is a block diagram showing an example of the configuration of a known decoding device.
  • FIG. 3 is a block diagram showing an example of the configuration of an encoding device to which the present invention is applied.
  • FIG. 4 is a block diagram showing an example of the configuration of a normalization coefficient encoder 51 .
  • FIG. 5 is an illustration of examples of normalization coefficients B.
  • FIG. 6 is a flowchart illustrating the operation of an encoder 61 - 1 .
  • FIG. 7 is an illustration of a probability distribution of differential values.
  • FIG. 8 is a flowchart illustrating the operation of an encoder 61 - 2 .
  • FIG. 9 is an illustration of the operation of the encoder 61 - 2 .
  • FIG. 10 is an illustration of a table A B .
  • FIG. 11 is an illustration of a table B B .
  • FIG. 12 is an illustration of a procedure for obtaining codes shown in the table B B .
  • FIG. 13 is an illustration of a table C B .
  • FIG. 14 is an illustration of a table D B .
  • FIG. 15 is a flowchart illustrating the details of step S 12 of FIG. 8 .
  • FIG. 16 is an illustration of a table E B .
  • FIG. 17 is another flowchart illustrating the details of step S 12 of FIG. 8 .
  • FIG. 18 is a flowchart illustrating the operation of an encoder 61 - 3 .
  • FIG. 19 is an illustration of the operation of the encoder 61 - 3 .
  • FIG. 20 is another illustration of a distribution of the differential values.
  • FIG. 21 is a flowchart illustrating the operation of an encoder 61 - 4 .
  • FIG. 22 is an illustration of the operation of the encoder 61 - 4 .
  • FIG. 23 is another illustration of a distribution of the differential values.
  • FIG. 24 is a flowchart illustrating the operation of an encoder 61 - 5 .
  • FIG. 25 is an illustration of the operation of the encoder 61 - 5 .
  • FIG. 26 is another illustration of a distribution of the differential values.
  • FIG. 27 is a flowchart illustrating the operation of an encoder 61 - 6 .
  • FIG. 28 is an illustration of the operation of the encoder 61 - 6 .
  • FIG. 29 is a block diagram showing an example of the configuration of a quantization accuracy information encoder 52 .
  • FIG. 30 is a flowchart illustrating the operation of an encoder 71 - 1 .
  • FIG. 31 is an illustration of an example of quantization accuracy information D.
  • FIG. 32 is an illustration of another distribution of the differential values.
  • FIG. 33 is a flowchart illustrating the operation of an encoder 71 - 2 .
  • FIG. 34 is an illustration of the operation of the encoder 71 - 2 .
  • FIG. 35 is an illustration of a table A D .
  • FIG. 36 is an illustration of a table B D .
  • FIG. 37 is an illustration of a table C D .
  • FIG. 38 is an illustration of another operation of the encoder 71 - 2 .
  • FIG. 39 is an illustration of another operation of the encoder 71 - 2 .
  • FIG. 40 is an illustration of another operation of the encoder 71 - 2 .
  • FIG. 41 is an illustration of another operation of the encoder 71 - 2 .
  • FIG. 42 is an illustration of the operation of an encoder 71 - 3 .
  • FIG. 43 is an illustration of another distribution of the differential values.
  • FIG. 44 is an illustration of the operation of an encoder 71 - 4 .
  • FIG. 45 is an illustration of another distribution of the differential values.
  • FIG. 46 is an illustration of the operation of an encoder 71 - 5 .
  • FIG. 47 is an illustration of another distribution of the differential values.
  • FIG. 48 is an illustration of the operation of an encoder 71 - 6 .
  • FIG. 49 is a block diagram showing an example of the configuration of a decoding device to which the present invention is applied.
  • FIG. 50 is a block diagram showing an example of the configuration of a normalization coefficient decoder 101 .
  • FIG. 51 is a block diagram showing an example of the configuration of a quantization accuracy information decoder 102 .
  • FIG. 52 is a flowchart illustrating the operation of the normalization coefficient decoder 101 .
  • FIG. 53 is a block diagram showing an example of the configuration of a computer 501 .
  • FIG. 3 shows an example of the configuration of an encoding device to which the present invention is applied.
  • a normalization coefficient encoder 51 and a quantization accuracy information encoder 52 are added to an encoding device shown in FIG. 1 . Since the remaining portion is similar to that shown in FIG. 1 , a description thereof is appropriately omitted.
  • normalizers 2 - 1 to 2 - 27 detect signal components, each having the largest absolute value, from signals of encoding unit A 0 to A 26 output from a band separator 1 , compute normalization coefficients B 0 to B 26 for the encoding units A 0 to A 26 using the detected values, and output the computed normalization coefficients to the normalization coefficient encoder 51 .
  • the normalizers 2 - 1 to 2 - 27 normalize the encoding units A 0 to A 26 in accordance with values corresponding to the computed normalization coefficients B 0 to B 26 , thus generating normalized data C 0 to C 26 , and output the normalized data C 0 to C 26 to quantizers 4 - 1 to 4 - 27 , respectively.
  • a quantization accuracy determination unit 3 determines quantization steps for quantizing the normalized data C 0 to C 26 in accordance with the encoding units A 0 to A 26 from the band separator 1 and outputs normalization accuracy information D 0 to D 26 , corresponding to the determined quantization steps, to the corresponding quantizers 4 - 1 to 4 - 27 and to the quantization accuracy information encoder 52 .
  • the normalization coefficient encoder 51 converts the normalization coefficients B 0 to B 26 from the normalizers 2 - 1 to 2 - 27 into corresponding predetermined indexes, encodes the indexes by various methods described below, and outputs resultant codes and information relating to the encoding methods to a multiplexer 5 .
  • the normalization coefficients B represent converted indexes.
  • the quantization accuracy information encoder 52 converts the quantization accuracy information D 0 to D 26 from the quantization accuracy determination unit 3 into corresponding predetermined indexes, encodes the indexes by various methods described below, and outputs the indexes to the multiplexer 5 .
  • the quantization accuracy information D represent converted indexes.
  • a band is separated into 27 sections, thus generating the 27 encoding units A.
  • the present invention is also applicable to cases in which more or fewer encoding units A are generated.
  • the normalization coefficient encoder 51 will now be described.
  • FIG. 4 shows an example of the configuration of the normalization coefficient encoder 51 .
  • the normalization coefficient encoder 51 has six encoders 61 - 1 to 61 - 6 .
  • the number of encoders 61 is not limited to six. It is only required that there be a plurality of encoders 61 .
  • the encoders 61 - 1 to 61 - 6 each encode the normalization coefficients B 0 to B 26 by methods described below and output the encoded normalization coefficients B 0 to B 26 to terminals connected to a switch 63 . Also, the encoders 61 -l to 61 - 6 each compute the total number of bits of codes obtained by encoding the normalization coefficients B 0 to B 26 and output the computation results to a determination unit 62 .
  • the determination unit 62 selects the encoder 61 that has output a minimum number of bits of the total number of bits of the codes from the encoders 61 - 1 to 61 - 6 and controls the switch 63 so that the code output by the selected encoder 61 can be output to the multiplexer 5 .
  • the determination unit 62 outputs information relating to an encoding method for the selected encoder 61 to the multiplexer 5 .
  • Each of the encoders 61 - 1 to 61 - 6 will now be described.
  • the normalization coefficients B 0 to B 26 in one frame, such as those shown in FIG. 5 are to be encoded will be described.
  • normalization coefficients do not change greatly at higher frequencies.
  • normalization coefficients for encoding units at higher frequencies often have similar values.
  • the encoder 61 - 1 subtracts, from the normalization coefficient for each encoding unit, a minimum value of the normalization coefficients of these encoding units. As a result, the encoder 61 - 1 encodes the normalization coefficients and outputs the encoded normalization coefficients. In this case, a normalization coefficient for an encoding unit below a predetermined band is not encoded and is output in its original form.
  • the operation of the encoder 61 - 1 is shown in a flowchart of FIG. 6 .
  • step S 1 the encoder 61 - 1 initializes the value of a built-in counter i B to 0, the value of a built-in register A B to 0, the value of a built-in register B B to the value X, which is sufficiently greater than a value computed in the following step S 4 , and the value of a built-in register C B to 0.
  • step S 3 the encoder 61 - 1 computes the difference between the maximum value and the minimum value detected in step S 2 and obtains the number of bits that can represent the computation result as the required number of bits.
  • step S 4 the encoder 61 - 1 computes the total number of bits by computing the sum of the number of bits of data representing the value of the counter i B , the number of bits of data representing the minimum value detected in step S 2 , the number of bits of data representing the required number of bits, which is obtained in step S 3 , the required number of bits ⁇ the number of the normalization coefficients B i to B 26 , and 6 bits ⁇ the number of the normalization coefficients B 0 to B i-1 .
  • the encoder 61 - 1 stores the computed total number of bits in the register A B (overwrites).
  • step S 5 the encoder 61 - 1 determines whether or not the value of the register A B is less than the value of the register B B . If it is determined that the value of the register A B is less than the value of the register B B , in step S 6 , the encoder 61 - 1 stores the value of the register A B in the register B B (overwrites) and stores the value of the counter i B at that time in the register C B (overwrites).
  • step S 7 the encoder 61 - 1 determines whether or not the value of the counter i B is 26. If it is determined that the value of the counter i B is not 26, in step S 8 , the encoder 61 - 1 increments the value of the counter i B by one and returns to step S 2 .
  • steps S 2 to S 8 is repeated until it is determined in step S 7 that the value of the counter i B is 26.
  • the minimum value of the total number of bits computed in step S 4 is stored in the register B B
  • the value of the counter i B at that time is stored in the register C B .
  • 2 is registered in the register C B .
  • step S 9 the encoder 61 - 1 subtracts the minimum value of the normalization coefficients B C to B 26 from each of the normalization coefficients B C to B 26 specified by the value of the register C B . As a result, the normalization coefficients B C to B 26 are encoded.
  • the register C B 2
  • the minimum value 37 is subtracted from each of the normalization coefficients B 2 to B 26 .
  • the normalization coefficients B 2 to B 26 are encoded.
  • step S 10 the encoder 61 - 1 outputs, to corresponding terminals connected to the switch 63 , the value of the register C B using data having a predetermined number of bits, the minimum value of the normalization coefficients B C to B 26 using data having a predetermined number of bits, the required number of bits at that time using data having a predetermined number of bits, each code of the normalization coefficients B C to B 26 computed in step S 9 using data having the required number of bits, and the normalization coefficients B 0 to B C-1 using data having a predetermined number of bits.
  • the value of the register C B that is, 2, is output using 5-bit data; the minimum value 37 is output using 6-bit data; the required number of bits, that is, 4 bits, is output using 3-bit data; each code of the normalization coefficients B 2 to B 26 is output using 4-bit data; and the normalization coefficients B 0 and B 1 (the normalization coefficients B (indexes) themselves) are each output using 6-bit data.
  • step S 11 the encoder 61 - 1 assumes the value of the register B B as the amount of encoding by the encoder 61 - 1 and outputs the value to the determination unit 62 .
  • normalization coefficients smoothly vary in the frequency direction, and the correlation in the frequency direction is high.
  • a particular biased distribution (a distribution in which a differential values near 0 have high probabilities of appearance), such as that shown in FIG. 7 , can be obtained from differential values between normalization coefficients corresponding to encoding units in adjacent bands.
  • IDSF denotes the index of a normalization coefficient. The same applies to other drawings.
  • the encoder 61 - 2 computes the differential values between normalization coefficients corresponding to encoding units in adjacent bands and allocates a code having a small number of bits to a differential value having a high probability of appearance. Accordingly, normalization coefficients are encoded.
  • the operation of the encoder 61 - 2 is shown in a flowchart of FIG. 8 .
  • step S 11 it is assumed that the encoding unit that corresponds to the normalization coefficient B i is denoted by A I .
  • step S 12 the encoder 61 - 2 refers to a table A B shown in FIG. 10 , reads a code corresponding to the differential value computed instep S 11 , and performs encoding.
  • computable differential values are coordinated with codes.
  • differential values for example, ⁇ 4 to 2
  • a code having a small number of bits is allocated to a normalization coefficient B corresponding to a differential value having a high probability of appearance.
  • the entirety of normalization coefficients B can be encoded by a small number of bits.
  • step S 13 the encoder 61 - 2 determines whether or not all the normalization coefficients B have been encoded. If it is determined that not all the normalization coefficients B have been encoded, the encoder 61 - 2 returns to step S 11 and performs similar processing on the next normalization coefficient.
  • step S 14 the encoder 61 - 2 outputs the code read in step S 12 to a terminal connected to the switch 63 , computes the total number of bits required to perform encoding, and outputs the computation result to the determination unit 62 . Subsequently, the processing is terminated.
  • a code of a normalization coefficient for a differential value other than differential values having a high probability of appearance can be written by, as shown in FIG. 12 , the 3 bits “100” (escape code) and a 6-bit index of the normalization coefficient to be encoded.
  • the code is written in the order of b 0 to b 8 .
  • the normalization coefficient is encoded into 9 bits.
  • the normalization coefficients B 0 to B 26 are encoded.
  • the total number of bits is, as indicated by the rightward arrow in FIG. 9 , 100 bits.
  • 62 bits can be saved.
  • a value corresponding to the “table B B ” in the drawing indicates the number of bits of codes of the normalization coefficients B 0 to B 26 , which are obtained using the table B B .
  • the total (100) is indicated by the arrow. The same applies to other drawings.
  • FIG. 13 shows a table C B in which codes corresponding to differential values having a high probability of appearance are changed.
  • the total number of bits of codes when the normalization coefficients B 0 to B 26 are encoded using the table C B is, as indicated by the rightward arrow in FIG. 9 , 96 bits. Accordingly, the total number of bits of codes can be adjusted by changing differential values with a high probability of appearance and corresponding codes.
  • a value corresponding to the “table C B ” in the drawing indicates the number of bits of codes of the normalization coefficients B 0 to B 26 , which are obtained using the table C B .
  • the total (96) is indicated by the arrow. The same applies to other drawings.
  • FIG. 14 shows a table D B in which another procedure for obtaining codes is coordinated with differential values other than differential values having a high probability of appearance.
  • a code of a normalization coefficient representing a differential value other than differential values having a high probability of appearance is regarded as data having an escape code representing the sign of the differential value (“1000” for the positive value and “1001” for the negative value), an escape code in accordance with the size of the differential value (“100”), and data of a size within a predetermined range based on the differential value.
  • This data is written in order from b 0 .
  • a code of a normalization coefficient representing a differential value other than differential values having a high probability of appearance can have a variable length.
  • the total number of bits of codes when the normalization coefficients B 0 to B 26 are encoded using the table D B is, as indicated by the rightward arrow in FIG. 9 , 92 bits.
  • a value corresponding to the “table D B ” in the drawing indicates the number of bits of codes of the normalization coefficients B 0 to B 26 , which are obtained using the table D B .
  • the total (92) is indicated by the arrow. The same applies to other drawings.
  • step S 12 ( FIG. 8 ) in which the encoder 61 - 2 uses the table D B to encode the normalization coefficients B will now be described with reference to a flowchart of FIG. 15 .
  • step S 21 the encoder 61 - 2 initializes the value of a counter N to 1.
  • step S 22 the encoder 61 - 2 determines whether or not the absolute value of the differential value computed in step S 11 ( FIG. 8 ) is greater than (3 ⁇ the value of the counter N). If it is determined that the absolute value is not greater than (3 ⁇ the value of the counter N), that is, if a differential value ( ⁇ 3 to 3) having a high probability of appearance is obtained, the encoder 61 - 2 proceeds to step S 23 .
  • step S 23 the encoder 61 - 2 refers to a table D and reads a code corresponding to the differential value having a high probability of appearance.
  • step S 22 if it is determined in step S 22 that the absolute value of the differential value is greater than (3 ⁇ the value of the counter N), that is, if a differential value other than that with a high profanity of appearance is obtained, the encoder 61 - 2 proceeds to step S 24 .
  • step S 24 the encoder 61 - 2 determines whether or not the differential value is a positive value. If it is determined that the differential value is a positive value, in step S 25 , the encoder 61 - 2 writes the 4 bits “1000” in order from b 0 of the bit configuration representing a code of the normalization coefficient B in the end (hereinafter referred to as the code bit configuration). In contrast, if it is determined in step S 24 that the differential value is not a positive value (if the differential value is a negative value), in step S 26 , the encoder 61 - 2 writes the 4 bits “1001” in order from b 0 of the code bit configuration.
  • step S 27 the encoder 61 - 2 increments the value of the counter N by one and proceeds to step S 28 .
  • step S 28 the encoder 61 - 2 determines whether or not the absolute value of the differential value is less than or equal to (3 ⁇ the value of the counter N+the value of the counter N ⁇ 1). If it is determined that the absolute value is not less than or equal to (3 ⁇ the value of the counter N+the value of the counter N ⁇ 1), in step S 29 , the encoder 61 - 2 writes the 3 bits “100” subsequent to the code bit configuration. Subsequently, the processing returns to step S 27 , and the processing from step S 27 onward is performed. Specifically, the 3 bits “100” is repeatedly written subsequent to the code bit configuration until it is determined in step S 28 that the differential value is less than or equal to (3 ⁇ the value of the counter N+the value of the counter N ⁇ 1).
  • step S 30 the encoder 61 - 2 determines whether or not the differential value is a positive value. If it is determined that the differential value is a positive value, the encoder 61 - 2 proceeds to step S 31 .
  • step S 31 the encoder 61 - 2 uses (the differential value ⁇ 4 ⁇ (the value of the counter N ⁇ 1)) as a new differential value and proceeds to step S 23 .
  • step S 23 a code corresponding to the new differential value is read from the table and written subsequent to the code bit configuration. Accordingly, the encoding of the normalization coefficient is terminated.
  • step S 30 if it is determined in step S 30 that the differential value is a negative value, the encoder 61 - 2 proceeds to step S 32 .
  • step S 32 the encoder 61 - 2 uses (the differential value +4 ⁇ (the value of the counter N ⁇ 1)) as a new differential value and proceeds to step S 23 .
  • step S 23 a code corresponding to the new differential value is read from the table and written subsequent to the code bit configuration. Accordingly, the encoding of the normalization coefficient is terminated.
  • step S 23 If the encoding of the normalization coefficient is terminated in step S 23 , the processing is terminated. Subsequently, the processing proceeds to step S 13 in FIG. 8 .
  • FIG. 16 shows another table E B .
  • the code of the normalization coefficient can have a variable-length or a fixed length depending on the size of the differential value.
  • the total number of bits of codes when the normalization coefficients B 0 to B 26 are encoded using the table E B is, as indicated by the rightward arrow in FIG. 9 , 95 bits.
  • a value corresponding to the “table E B ” in the drawing indicates the number of bits of codes of the normalization coefficients B 0 to B 26 , which are obtained using the table E B .
  • the total (95) is indicated by the arrow. The same applies to other drawings.
  • step S 12 ( FIG. 8 ) in which the normalization coefficients B are encoded using the table E B will now be described with reference to a flowchart of FIG. 17 .
  • step S 41 the encoder 61 - 2 determines whether or not the absolute value of the differential value computed in step S 11 ( FIG. 8 ) is greater than 3. If it is determined that the absolute value is not greater than 3, that is, if a differential value ( ⁇ 3 to 3) having a high probability of appearance is obtained, the encoder 61 - 2 proceeds to step S 42 .
  • step S 42 the encoder 61 - 2 refer to a table E and reads codes corresponding to differential values having a high probability of appearance.
  • step S 41 determines whether or not the absolute value of the differential value is less than or equal to 7. If it is determined that the absolute value is not less than or equal to 7 (greater than 7), the encoder 61 - 2 proceeds to step S 44 .
  • step S 44 the encoder 61 - 2 writes the 4 bits “1001” and the 6-bit normalization coefficient value in order from b 0 of the bit configuration (code bit configuration) representing the code of the normalization coefficient in the end. Accordingly, the encoding of the normalization coefficient is terminated.
  • step S 45 the encoder 61 - 2 writes the 4 bits “1000” in order from b 0 of the code bit configuration.
  • step S 46 the encoder 61 - 2 determines whether or not the differential value is a positive value. If it is determined that the differential value is a positive value, in step S 47 , the encoder 61 - 2 regards (the differential value ⁇ 4) as a new differential value and proceeds to step S 42 .
  • step S 46 determines whether the differential value is a negative value.
  • step S 48 the encoder 61 - 2 uses (the differential value +3) as a new differential value and proceeds to step S 42 .
  • step S 42 When a code is read from the table in step S 42 and the encoding of the normalization coefficient is completed, the processing is terminated. Subsequently, the encoder 61 - 2 proceeds to step S 13 in FIG. 8 .
  • the normalization coefficients vary smoothly in the frequency direction. In many cases, the values decrease from lower to higher frequencies. In other words, by adding a weight value which increases step-by-step from lower to higher frequencies to each normalization coefficient, the differential value between normalization coefficients corresponding to encoding units in adjacent bands can be reduced, and hence the frequency of appearance of a differential value having a high frequency of occurrence of appearance can be further increased.
  • the encoder 61 - 3 adds a weight value which increases step-by-step from lower to higher frequencies to each of the normalization coefficients B.
  • the encoder 61 - 3 computes the differential value between the normalization coefficients B, corresponding to encoding units in adjacent bands, to which the weight values are added (hereinafter referred to as weighted normalization coefficients BW).
  • the encoder 61 - 3 allocates a code having a small number of bits to a differential value having a high probability of appearance to encode the normalization coefficients B.
  • the operation of the encoder 61 - 3 is shown in a flowchart of FIG. 18 .
  • step S 51 the encoder 61 - 3 computes a weight value to be added to each normalization coefficient B. Specifically, for example, (i/slope of weight curve (3 in this example)) is computed, and an integer portion thereof is used as the weight value for the normalization coefficient B i . In other words, as shown in FIG.
  • the weight value for the normalization coefficients B 0 to B 2 is 0; the weight value for the normalization coefficients B 3 to B 5 is 1; the weight value for the normalization coefficients B 6 to B 8 is 2; the weight value for the normalization coefficients B 9 to B 11 is 3; the weight value for the normalization coefficients B 12 to B 14 is 4; the weight value for the normalization coefficients B 15 to B 17 is 5; the weight value for the normalization coefficients B 18 to B 20 is 6; the weight value for the normalization coefficients B 21 to B 23 is 7; and the weight value for the normalization coefficients B 24 to B 26 is 8.
  • the index values shown in FIG. 19 are the same as those shown in FIG. 5 .
  • step S 52 the encoder 61 - 3 adds the weight values computed in step S 51 to the corresponding normalization coefficients B, thus obtaining weighted normalization coefficients BW.
  • step S 53 the encoder 61 - 3 computes the differential value between the weighted normalization coefficient BW i obtained from the normalization coefficient B i to be encoded and the weighted normalization coefficient BW i-1 obtained from the normalization coefficient B i-1 corresponding to the encoding unit A i-1 adjacent to the lower side of the encoding unit A i corresponding to the normalization coefficient B i .
  • steps from S 54 to S 56 processing similar to that in steps from S 12 to S 14 of FIG. 8 is performed, and a description thereof is omitted.
  • the total numbers of bits of codes obtained in cases in which the encoder 61 - 3 encodes the normalization coefficients B 0 to B 26 using the table B B ( FIG. 11 ), the table C B ( FIG. 13 ), the table D B ( FIG. 14 ), and the table E B ( FIG. 16 ) are, as indicated by the rightward arrows in FIG. 19 , 104 bits, 90 bits, 89 bits, and 91 bits, respectively.
  • a known case (162 bits) 58 bits, 72 bits, 73 bits, or 71 bits can be saved.
  • bits can be saved except for a case in which the table B B is used.
  • the above-described encoder 61 - 1 can compute weight values by processing corresponding to step S 51 described above and can compute weighted normalization coefficients BW by processing corresponding to step S 52 . In other words, in this case, processing in steps S 1 to S 11 is performed on the weighted normalization coefficients BW.
  • a particular biased distribution (having a high probability near 0), such as that shown in FIG. 20 , can be obtained from differential values between normalization coefficients that are temporarily adjacent to each other.
  • the encoder 61 - 4 computes the differential value between normalization coefficients that are temporarily adjacent to each other and allocates a code having a small number of bits to a differential value having a high probability of appearance to encode normalization coefficients
  • the operation of the encoder 61 - 4 is shown in a flowchart of FIG. 21 .
  • step S 61 the encoder 61 - 4 computes the differential value between the normalization coefficient B i of the current frame and the normalization coefficient B i of a frame one frame temporarily prior to the current frame.
  • the normalization integers B of the current frame shown in FIG. 22 are the same as those shown in FIG. 5 .
  • step S 62 to S 64 processing similar to that in steps S 12 to S 14 of FIG. 8 is performed, and a description thereof is omitted.
  • the total numbers of bits of codes obtained in cases in which the encoder 61 - 4 encodes the normalization coefficients B 0 to B 26 using the table B B , the table C B , the table D B , and the table E B are, as indicated by the rightward arrows in FIG. 22 , 96 bits, 105 bits, 93 bits, and 98 bits, respectively.
  • left-signal and right-signal channels of a stereo audio signal have similar power (the correlation between channels is high).
  • normalization coefficients B L of a left signal and normalization coefficients B R of a right signal are close to each other.
  • the differential values between the normalization coefficients B L of a left signal and the normalization coefficients B R of a right signal have a particular biased distribution (a distribution in which the differential values near 0 have high probabilities).
  • the encoder 61 - 5 computes the differential values between the normalization coefficients B L of a corresponding left signal and the normalization coefficients B R of a corresponding right signal and allocates a code having a small number of bits to a differential value having a high probability of appearance to encode the normalization coefficients B.
  • normalization coefficients B L0 to B L26 of encoding units A L0 to A L26 generated by separating a left signal into 27 bands and normalization coefficients B R0 to B R26 of encoding units A R0 to A R26 generated by separating a right signal into 27 bands are input to the encoder 61 - 5 .
  • the operation of the encoder 61 - 5 is shown in a flowchart of FIG. 24 .
  • step S 71 the encoder 61 - 5 computes the corresponding differential values between the normalization coefficients B Li and the normalization coefficients B Ri to be encoded.
  • steps S 72 to S 74 processing similar to that in steps S 12 to S 14 of FIG. 8 is performed, and a detailed description thereof is omitted.
  • the total numbers of bits of codes obtained in cases in which the encoder 61 - 5 encodes the normalization coefficients B L0 to B L26 and the normalization coefficients B R0 to B R26 using the table B B , the table C B , the table D B , and the table E B are, as indicated by the rightward arrows in FIG. 25 , 96 bits, 91 bits, 84 bits, and 84 bits, respectively.
  • normalization coefficients are highly correlated between adjacent bands (frequency direction) and channels.
  • the differential values have, as shown in FIG. 26 , a particular biased distribution (a distribution in which the differential values near 0 have high probabilities).
  • the encoder 61 - 6 computes the differential values LR between the normalization coefficients B L of a corresponding left signal and the normalization coefficients B R of a corresponding right signal. Subsequently, the encoder 61 - 6 computes the differential values F in the frequency direction between these differential values LR. The encoder 61 - 6 allocates a code having a small number of bits to the differential value F having a high probability of appearance to encode the normalization coefficients B.
  • the operation of the encoder 61 - 6 is shown in a flowchart of FIG. 27 .
  • step S 81 the encoder 61 - 6 computes the differential values LR between the corresponding normalization coefficients B L and the normalization coefficients B R .
  • step S 82 the encoder 61 - 6 computes the differential value F between the differential value LR between the normalization coefficient B Li and the normalization coefficient B Ri to be encoded and the differential value LR between the normalization coefficient B Li-1 and the normalization coefficient B Ri-1 .
  • the normalization coefficients B Li-1 and the normalization coefficient B Ri-1 correspond to encoding units A Li-1 and A Ri-1 adjacent to the lower side of an encoding unit A Li and an encoding unit A Ri corresponding to the normalization coefficient B Li and the normalization coefficient B Ri .
  • steps S 83 to S 85 processing similar to that in steps S 12 to S 14 of FIG. 8 is performed, and a description thereof is omitted.
  • the total numbers of bits of codes obtained in cases in which the encoder 61 - 6 encodes the normalization coefficients B 0 to B 26 using the table B B , the table C B , the table D B , and the table E B are, as indicated by the rightward arrows in FIG. 28 , 114 bits, 90 bits, 89 bits, and 91 bits, respectively.
  • the encoders 61 - 1 to 61 - 6 of the normalization coefficient encoder 51 encode normalization coefficients using predetermined encoding methods.
  • the quantization accuracy information encoder 52 will now be described.
  • FIG. 29 shows an example of the configuration of the quantization accuracy information encoder 52 .
  • the quantization accuracy information encoder 52 has six encoders 71 - 1 to 71 - 6 .
  • the number of encoders 71 is not limited to six. It is only required that there be a plurality of encoders 71 .
  • the encoders 71 - 1 to 71 - 6 each encode quantization accuracy information D 0 to D 26 by methods described below and output the encoded quantization accuracy information D 0 to D 26 to terminals connected to a switch 73 . Also, the encoders 71 - 1 to 71 - 6 each compute the total number of bits of codes obtained by encoding the quantization accuracy information D 0 to D 26 and output the computation results to a determination unit 72 .
  • the determination unit 72 selects the encoder 71 that has output a minimum number of bits of the total number of bits of the codes from the encoders 71 - 1 to 71 - 7 and controls the switch 73 so that the code output by the selected encoder 71 can be output to the multiplexer 5 .
  • the determination unit 72 outputs information relating to an encoding method for the selected encoder 71 to the multiplexer 5 .
  • quantization accuracy information values do not change greatly at higher frequencies.
  • pieces of quantization accuracy information for encoding units at higher frequencies often have similar values.
  • the encoder 71 - 1 subtracts, from the quantization accuracy information for each encoding unit, a minimum value of the quantization accuracy information of these encoding units. As a result, the encoder 71 - 1 encodes the quantization accuracy information and outputs the encoded quantization accuracy information. In this case, quantization accuracy information for an encoding unit below a predetermined band is not encoded and is output in its original form.
  • the operation of the encoder 71 - 1 is shown in a flowchart of FIG. 30 .
  • step S 91 the encoder 71 - 1 initializes the value of a built-in counter i D to 0, the value of a built-in register A D to 0, the value of a built-in register B D to the value Y, which is sufficiently greater than a value computed in the following step S 94 , and the value of a built-in register C D to 0.
  • step S 93 the encoder 71 - 1 computes the difference between the maximum value and the minimum value detected in step S 92 and obtains the number of bits that can represent the computation result as the required number of bits.
  • step S 94 the encoder 71 - 1 computes the total number of bits by computing the sum of the number of bits of data representing the value of the counter i D , the number of bits of data representing the minimum value detected in step S 92 , the number of bits of data representing the required number of bits, which is obtained in step S 93 , the required number of bits ⁇ the number of pieces of the quantization accuracy information D i to D 26 , and 3 bits ⁇ the number of pieces of the quantization accuracy information D 0 to D i-1 .
  • the encoder 71 - 1 stores the computed total number of bits in the register A D (overwrites).
  • the counter i D 14, 65 bits, which is the sum of 5 bits of data representing the value 14 of the counter i D , 3 bits of data representing the minimum value 1, 2 bits of data representing 1 bit, i.e., the required number of bits, the required number of bits (2 bits) ⁇ 13 (the number of pieces of the quantization accuracy information D 14 to D 26 ), and 3 bits ⁇ 14 (the number of pieces of the quantization accuracy information D 0 and D 13 ), is regarded as the total number of bits and is stored in the register A D .
  • step S 95 the encoder 71 - 1 determines whether or not the value of the register A D is less than the value of the register B D . If it is determined that the value of the register A D is less than the value of the register B D , in step S 96 , the encoder 71 - 1 stores the value of the register A D in the register B D (overwrites) and stores the value of the counter i D at that time in the register C D (overwrites).
  • step S 97 the encoder 71 - 1 determines whether or not the value of the counter i D is 26. If it is determined that the value of the counter i D is not 26, in step S 98 , the encoder 71 - 1 increments the value of the counter i D by one and returns to step S 92 .
  • steps S 92 to S 98 is repeated until it is determined in step S 97 that the value of the counter i D is 26.
  • the minimum value of the total number of bits computed in step S 94 is stored in the register B D
  • the value of the counter i D at that time is stored in the register C D .
  • the total number of bits (65) computed when the counter i D 14 is registered in the register B D
  • 14 is registered in the register C D .
  • step S 99 the encoder 71 - 1 subtracts the minimum value of the quantization accuracy information D C to D 26 from each of the quantization accuracy information D C to D 26 specified by the value of the register C D . As a result, the pieces of the quantization accuracy information D C to D 26 are encoded.
  • the minimum value 1 is subtracted from each of the quantization accuracy information D 14 to D 26 .
  • the pieces of the quantization accuracy information D 14 to D 26 are encoded.
  • step S 110 the encoder 71 - 1 outputs, to corresponding terminals connected to the switch 73 , the value of the register C D using data having a predetermined number of bits, the minimum value of the quantization accuracy information D C to D 26 using data having a predetermined number of bits, the required number of bits at that time using data having a predetermined number of bits, each code of the quantization accuracy information D C to D 26 computed in step S 99 using data having the required number of bits, and the quantization accuracy information D 0 to D C-1 using data having a predetermined number of bits.
  • the value of the register C D that is, 14, is output using 5-bit data; the minimum value 1 is output using 3-bit data; the required number of bits, that is, 1 bit, is output using 2-bit data; each code of the quantization accuracy information D 14 to D 26 is output using 1-bit data; and the quantization accuracy information D 0 to D 13 (the quantization accuracy information D (indexes) themselves) are each output using 3-bit data.
  • step S 101 the encoder 71 - 1 assumes the value of the register B D as the amount of encoding by the encoder 71 - 1 and outputs the value to the determination unit 72 .
  • quantization accuracy information smoothly varies in the frequency direction, and the correlation in the frequency direction is high.
  • a particular biased distribution such as that shown in FIG. 32 , can be obtained from differential values between pieces of quantization accuracy information corresponding to encoding units in adjacent bands.
  • the encoder 71 - 2 computes the differential values between quantization accuracy information corresponding to encoding units in adjacent bands and allocates a code having a small number of bits to a differential value having a high probability of appearance. Accordingly, quantization accuracy information is encoded.
  • the operation of the encoder 71 - 2 is shown in a flowchart of FIG. 33 .
  • step S 111 the encoder 71 - 2 computes the differential value between quantization accuracy information D i to be encoded and quantization accuracy information D i-1 .
  • step S 112 the encoder 71 - 2 determines codes for the quantization accuracy information D i using a table A D ( FIG. 35 ), a table B D ( FIG. 36 ), or a table C D ( FIG. 37 ). Since the method for encoding the quantization accuracy information D using the table A D , the table B D , or the table C D is basically the same as that for encoding the normalization coefficients B described with reference to FIG. 8 , a description thereof is omitted.
  • step S 113 the encoder 71 - 2 determines whether or not all the pieces of quantization accuracy information D have been encoded. If it is determined that not all the pieces of quantization accuracy information D have been encoded, the encoder 71 - 2 returns to step S 111 and performs similar processing on the next piece of quantization accuracy information D.
  • step S 114 the encoder 71 - 2 outputs the code determined in step S 112 to a terminal connected to the switch 73 , computes the total number of bits, and outputs the computation result to the determination unit 72 . Subsequently, the processing is terminated.
  • the total numbers of bits of codes obtained in cases in which the encoder 71 - 2 encodes the quantization accuracy information D 0 to D 26 using the tables A D to C D are, as indicated by the rightward arrows in FIG. 34 , 49 bits, 39 bits, and 49 bits, respectively. In other words, compared with a known case (81 bits), 32 bits, 42 bits, or 32 bits can be saved.
  • pieces of quantization accuracy information D corresponding to encoding units at higher frequencies tend to indicate 0 or 1.
  • the encoder 71 - 2 may not encode these pieces of quantization accuracy information D.
  • the quantization accuracy information D 16 to D 26 are not encoded.
  • the total number of bits is 39 bits (the value indicated by the rightward arrow corresponding to the “table B D (no cut)” in the drawing).
  • the bits to be used to encode the quantization accuracy information can be saved further. It is regarded that 5 bits is the number of bits of information (such as 16) indicating the quantization accuracy information D 16 in the lowest band of the corresponding encoding unit of the series of the quantization accuracy information D 16 to D 26 indicating 0 or 1.
  • the quantization accuracy information D 18 to D 26 are not encoded, the total number of bits is 35 bits.
  • Is is regarded that 23 bits is the total number of bits of codes in a case in which the quantization accuracy information D 0 to D 13 are encoded using the table B D ; 5 bits is the number of bits of information (for example, 14) indicating information (for example, 14) indicating the quantization accuracy information D 14 in the lowest band of the corresponding unit A of the continuous series of the mixed pieces of quantization accuracy information D 14 to D 26 indicating 0 or 1; and 13 is the total number of bits of each code (1 bit) for the quantization accuracy information D 14 to D 26 .
  • a continuous series of mixed pieces of quantization accuracy information D 8 to D 26 indicating 0 to 3 can be encoded by 2 bits.
  • a continuous series of pieces of quantization accuracy information indicating 0 or 1 may not be encoded, or a continuous series of pieces of quantization accuracy information indicating a predetermined value may be encoded by a predetermined number of bits.
  • a continuous series of pieces of normalization coefficients information indicating 0 or 1 may not be encoded, or a continuous series of pieces of normalization coefficients information indicating a predetermined value may be encoded by a predetermined number of bits.
  • quantization accuracy information varies smoothly in the frequency direction. If the number of bits to be encoded is small, quantization accuracy information often decreases from low to high frequencies. Furthermore, in many cases, quantization accuracy information of the encoding unit A at low frequencies has a high value (index) such as 5, 6, or 7. In contrast, in many cases, quantization accuracy information D of an encoding unit at high frequencies has a low value (index) such as 0, 1, or 2.
  • the encoder 71 - 3 subtracts a weighting curve with a predetermined slope (coefficient) from each piece of quantization accuracy information, thereby reducing the differential value between quantization accuracy information of encoding units in adjacent bands and increasing the probability of appearance of the differential value having a high probability of appearance. As a result, the quantization accuracy information can be encoded using a smaller number of bits.
  • a weight value for quantization accuracy information D 0 and D 1 is 5, 5 is subtracted from the quantization accuracy information D 0 and D 1 .
  • a weight value for quantization accuracy information D 2 and D 3 is 4, 4 is subtracted from the quantization accuracy information D 2 and D 3 .
  • a weight value for quantization accuracy information D 4 and D 5 is 3, 3 is subtracted from the quantization accuracy information D 4 and D 5 .
  • a weight value for quantization accuracy information D 6 and D 7 is 2, 2 is subtracted from the quantization accuracy information D 6 and D 7 .
  • a weight value for quantization accuracy information D 8 and D 9 is 1, 1 is subtracted from the quantization accuracy information D 8 and D 9 . Since a weight value for quantization accuracy information D 10 to D 26 is 0, the quantization accuracy information D 10 to D 26 are unaltered.
  • the total numbers of bits of codes in cases in which the encoder 71 - 3 encodes the quantization accuracy information D 0 to D 26 using the table A D ( FIG. 35 ) and the table B D ( FIG. 36 ) are, as indicated by the rightward arrows in FIG. 42 , 47 bits and 38 bits, respectively. Compared with a known case (81 bits), 34 bits or 43 bits can be saved.
  • a continuous series of pieces of quantization accuracy information having 0 or 1 may not be encoded, or a continuous series of pieces of quantization accuracy information having a predetermined value may be encoded by a predetermined number of bits.
  • the above-described encoder 71 - 1 can compute weight values and can subtract the weight values from corresponding pieces of quantization accuracy information. In this case, processing in steps S 91 to S 101 is performed on the quantization accuracy information from which the weight values are subtracted.
  • a particular biased distribution (a distribution in which differential values near 0 have high probabilities), such as that shown in FIG. 43 , can be obtained from differential values between pieces of quantization accuracy information, which are temporarily adjacent to each other.
  • the encoder 71 - 4 computes the differential value between pieces of quantization accuracy information, which are temporarily adjacent to each other, and allocates a code having a small number of bits to a differential value having a high probability of appearance to encode the quantization accuracy information.
  • the operation of the encoder 71 - 4 is basically the same as that of the encoder 61 - 4 of the normalization coefficient encoder 51 , a detailed description thereof is omitted.
  • the total number of bits of codes in a case in which the encoder 71 - 4 encodes the quantization accuracy information D 0 to D 26 using the table A D is, as indicated by the rightward arrow in FIG. 44 , 45 bits. Compared with a known case (81 bits), 36 bits can be saved.
  • a continuous series of pieces of quantization accuracy information having 0 or 1 may not be encoded, or a continuous series of quantization accuracy information having a predetermined number may be encoded by a predetermined number of bits.
  • the operation of the encoder 71 - 5 is basically the same as that of the encoder 61 - 5 of the normalization coefficient encoder 51 , a detailed description thereof is omitted.
  • the total number of bits of codes in a case in which the encoder 71 - 5 encodes the quantization accuracy information D 0 to D 26 using the table A D is, as indicated by the rightward arrow in FIG. 46 , 49 bits. Compared with a known case (81 bits), 32 bits can be saved.
  • a continuous series of pieces of quantization accuracy information having 0 or 1 may not be encoded, or a continuous series of quantization accuracy information having a predetermined number maybe encoded by a predetermined number of bits.
  • pieces of quantization accuracy information D are highly correlated between adjacent bands (frequency direction) and channels.
  • differential values in the frequency direction are computed from differential values between channels, the differential values have, as shown in FIG. 47 , a particular biased distribution (a distribution in which the differential values near 0 have high probabilities).
  • the encoder 71 - 6 computes the differential values LR between the quantization accuracy information D L of a corresponding left signal and a right signal D R . Subsequently, the encoder 71 - 6 computes the differential values F in the frequency direction between these differential values LR. The encoder 71 - 6 allocates a code having a small number of bits to the differential value F having a high probability of appearance to encode the quantization accuracy information D.
  • the operation of the encoder 71 - 6 is basically the same as that of the encoder 61 - 6 of the normalization coefficient encoder 51 , a detailed description thereof is omitted.
  • the total number of bits of codes in a case in which the encoder 71 - 6 encodes the quantization accuracy information D 0 to D 26 using the table A is, as indicated by the rightward arrow in FIG. 48 , 59 bits. Compared with a known case, 22 bits can be saved.
  • a continuous series of pieces of quantization accuracy information having 0 or 1 may not be encoded, or a continuous series of quantization accuracy information having a predetermined number may be encoded by a predetermined number of bits.
  • FIG. 49 shows an example of the configuration of a decoding device to which the present invention is applied.
  • a normalization coefficient decoder 101 and a quantization accuracy information decoder 102 are added to a decoding device shown in FIG. 2 . Since the remaining portion is similar to that shown in FIG. 2 , a description thereof is appropriately omitted.
  • a demultiplexer 21 decodes encoded data into codes generated by encoding the normalization coefficients B 0 to B 26 (hereinafter referred to as encoded normalization coefficients U B ), information relating to methods for encoding the normalization coefficients B (hereinafter referred to as encoding information W B ), codes generated by encoding the quantization accuracy information D 0 to D 26 (hereinafter referred to as encoded quantization accuracy information U D ), information relating to methods for encoding the quantization accuracy information D (hereinafter referred to as encoding information W D ), and the quantization coefficients F 0 to F 26 .
  • the demultiplexer 21 outputs the encoded normalization coefficients U B and the encoding information W B to the normalization coefficient decoder 101 and outputs the encoded quantization accuracy information U D and the encoding information W D to the quantization accuracy information decoder 102 .
  • the demultiplexer 21 outputs the quantization coefficients F 0 to F 26 to corresponding signal component composers 22 - 1 to 22 - 27 .
  • the normalization coefficient decoder 101 decodes the encoded normalization coefficients U B from the demultiplexer 21 by a decoding method corresponding to the decoding information W B and outputs the resultant normalization coefficients B 0 to B 26 to the corresponding signal component composers 22 - 1 to 22 - 26 .
  • the quantization accuracy information decoder 102 decodes the encoded quantization accuracy information U D from the demultiplexer 21 by a decoding method corresponding to the encoding information W D and outputs the resultant quantization accuracy information D 0 to D 26 to the corresponding signal component composers 22 - 1 to 22 - 27 .
  • the signal component composers 22 - 1 to 22 - 27 dequantize the quantization coefficients F 0 to F 26 from the demultiplexer 21 in accordance with quantization steps corresponding to the quantization accuracy information D 0 to D 26 from the quantization accuracy information decoder 102 , multiply the resultant normalized data C 0 to C 26 by values corresponding to the normalization coefficients B 0 to B 26 from the normalization coefficient decoder 101 , thereby decoding signals of encoding units in each band.
  • the signals are output to a band combiner 23 .
  • FIG. 50 shows an example of the configuration of the normalization coefficient decoder 101 .
  • the normalization coefficient decoder 101 has six decoders 111 - 1 to 111 - 6 .
  • the number of decoders 111 is not limited to six. It is only required that there be a plurality of decoders 111 .
  • the encoded normalization coefficients U B are appropriately input to the decoders 111 - 1 to 111 - 6 .
  • the decoders 111 - 1 to 111 - 6 each perform decoding processing corresponding to the encoding processing by the encoders 61 - 1 to 61 - 6 of the normalization coefficient encoder 51 of the encoding device and each output the decoding results to terminals connected to a switch 114 .
  • a determination unit 112 determines the encoding contents of the encoded normalization coefficients U B (for example, which decoder of the decoders 61 - 1 to 61 - 6 has encoded the normalization coefficients and which table is used to encode the normalization coefficients) and, based on the determination results, selects a decoder to perform decoding processing and a table to be used to perform decoding.
  • the determination unit 112 controls a switch 113 so that the encoded normalization coefficients U B can be input to the selected decoder (any of 111 - 1 to 111 - 6 ) and controls the switch 114 so that the output of the selected decoder can be output to the signal component composers 22 .
  • FIG. 51 shows an example of the configuration of the quantization accuracy information decoder 102 .
  • the quantization accuracy information decoder 102 has six decoders 121 - 1 to 121 - 6 .
  • the number of decoders 121 is not limited to six. It is only required that there be a plurality of decoders 121 .
  • the encoded quantization accuracy information U D is appropriately input to the decoders 121 - 1 to 121 - 6 .
  • the decoders 121 - 1 to 121 - 6 each perform decoding processing corresponding to the encoding processing by the encoders 71 - 1 to 71 - 6 of the quantization accuracy information encoder 52 of the encoding device and each output the decoding results to terminals connected to a switch 124 .
  • a determination unit 122 determines the encoding contents of the encoded quantization accuracy information U D (for example, which one of the decoders 71 has encoded the quantization accuracy information and which table is used to encode the quantization accuracy information) and, based on the determination results, selects a decoder to perform decoding processing and a table to be used to perform decoding.
  • the determination unit 112 controls a switch 123 so that the encoded quantization accuracy information U D can be input to the selected decoder and controls the switch 124 so that the output of the selected decoder can be output to the signal component composers 22 .
  • step S 201 the normalization coefficient decoder 101 reads the first four bits of the encoded normalization coefficients U B .
  • step S 204 since the differential value 0 is encoded by 2 bits, the normalization coefficient decoder 101 moves the data reading start position backwards by 2 bits, which is obtained by subtracting 2 bits from 4 bits.
  • step S 202 determines whether or not the data corresponds to any of “0100”, to “0101”. If it is determined that the data corresponds to any of “0100” to “0101”, the normalization coefficient decoder 111 proceeds to step S 206 .
  • step S 207 since the differential value ⁇ 1 is encoded by 3 bits, the normalization coefficient decoder 101 moves the data reading start position backwards by 1 bit, which is obtained by subtracting 3 bits from 4 bits.
  • step S 208 the normalization coefficient decoder 101 determines whether or not the data corresponds to any of “0110” to “0111”. If it is determined that the data corresponds to any of “0110” to “0111”, the normalization coefficient decoder 101 proceeds to step S 209 .
  • step S 210 since the differential value ⁇ 2 is encoded by 3 bits, the normalization coefficient decoder 101 moves the data reading start position backwards by 1 bit, which is obtained by subtracting 3 bits from 4 bits.
  • step S 208 determines whether or not the data corresponds to any of “1000” to “1001”. If it is determined that the data corresponds to any of “1000” to “1001”, the normalization coefficient decoder 101 proceeds to step S 212 .
  • step S 212 the normalization coefficient decoder 101 determines that the escape code “101” (3 bits) in the table B B is used, moves the data reading start position backwards by 1 bit, which is obtained by subtracting 3 bits from four bits, and proceeds to step S 213 .
  • step S 213 the normalization coefficient decoder 101 reads 6-bit data from a bit at the moved reading start position (the last bit of the 4 bits read in step S 201 ) as the normalization coefficients B.
  • step S 211 If it is determined in step S 211 that the data written on the 4 bits corresponds to none of “1000” to “1001”, the normalization coefficient decoder 101 proceeds to step S 214 .
  • step S 214 the normalization coefficient decoder 101 determines whether or not the data corresponds to any of “1010” to “1011”. If it is determined that the data corresponds to any of “1010” to “1011”, the normalization coefficient decoder 101 proceeds to step S 215 .
  • step S 216 since the differential value 1 is encoded by 3 bits, the normalization coefficient decoder 101 moves the data reading start position backwards by 1 bit, which is obtained by subtracting 3 bits from 4 bits.
  • step S 217 the normalization coefficient decoder 101 determines whether or not the data corresponds to any of “1100” to “1101”. If it is determined that the data corresponds to any of “1100” to “1101”, the normalization coefficient decoder 101 proceeds to step S 218 .
  • step S 219 since the differential value ⁇ 3 is encoded by 3 bits, the normalization coefficient decoder 101 moves the data reading start position backwards by 1 bit, which is obtained by subtracting 3 bits from 4 bits.
  • step S 217 If it is determined in step S 217 that data written on the 4 bits corresponds to none of “1100” to “1101”, in step S 220 , the normalization coefficient decoder 101 determines whether or not the data corresponds to “1110”. If it is determined that the data corresponds to “1110”, the normalization coefficient decoder 101 proceeds to step S 221 .
  • step S 220 If it is determined in step S 220 that the data written on the 4 bits does not correspond to “1110”, the normalization coefficient decoder 101 proceeds to step S 223 .
  • step S 204 When the processing in step S 204 , step S 207 , step S 210 , step S 213 , step S 216 , step S 219 , step S 221 , or step S 223 is performed, the processing is terminated.
  • the present invention can be applied to an audio recorder/player.
  • a series of the above-descried processes can be performed by hardware or by software.
  • a program forming the software is installed in a computer, and the program is run on the computer.
  • the functions of the foregoing encoding device and the decoding device are realized.
  • FIG. 53 is a block diagram showing the configuration of an embodiment of a computer 501 functioning as the foregoing encoding device and the decoding device.
  • An input/output (I/O) interface 516 is connected through a bus 515 to a CPU (Central Processing Unit) 511 .
  • CPU Central Processing Unit
  • the CPU 511 If a command is input through the I/O interface 516 by a user from an input unit 518 including a keyboard and a mouse, the CPU 511 loads a program stored in a recording medium such as a ROM (Read Only Memory) 512 , a hard disk 514 , or a magnetic disk 531 , an optical disk 532 , a magneto-optical disk 533 , or a semiconductor memory 534 mounted on a drive 520 into a RAM (Random Access Memory) 513 and executes the program. As a result, the above-described various processes can be performed. Also, the CPU 511 outputs the processing results through the I/O interface 516 to an output unit 517 including an LCD (Liquid Crystal Display) if necessary.
  • a recording medium such as a ROM (Read Only Memory) 512 , a hard disk 514 , or a magnetic disk 531 , an optical disk 532 , a magneto-optical disk 533 , or a semiconductor memory 5
  • the program can be stored beforehand in the hard disk 514 or the ROM 512 and can be provided together with the computer 501 to the user. Also, the program can be provided as a packaged media such as the magnetic disk 531 , the optical disk 532 , the magneto-optical disk 533 , or the semiconductor memory 534 . It is also possible to provide the program from a satellite or a network to the hard disk 514 through a communication unit 519 .
  • steps for writing a program provided by a recording medium not only include time-series processing performed in accordance with the described order but also include parallel or individual processing, which may not necessarily be performed in time series.
  • an input acoustic time-series signal is separated into N bands to generate N band signals.
  • a predetermined normalization coefficient for each of the band signals is generated.
  • the band signals are normalized on the basis of the generated normalization coefficients.
  • Quantization accuracy information for each of the band signals is generated.
  • Each of the normalized band signals is quantized on the basis of the quantization accuracy information.
  • the N generated normalization coefficients are encoded.
  • One of normalization coefficient encoding methods is selected on the basis of the amount of encoding in a case in which the N normalization coefficients are encoded.
  • the N normalization coefficients, which are encoded using the selected normalization coefficient encoding method, and each of the band signals are multiplexed. Accordingly, the encoding efficiency can be improved.
  • an input acoustic time-series signal is separated into N bands to generate N band signals.
  • a predetermined normalization coefficient for each of the band signals is generated.
  • the band signals are normalized on the basis of the generated normalization coefficients.
  • Quantization accuracy information for each of the band signals is generated.
  • Each of the normalized band signals is quantized on the basis of the quantization accuracy information.
  • the N pieces of generated quantization accuracy information are encoded.
  • One of quantization accuracy information encoding methods is selected on the basis of the amount of encoding in a case in which the N pieces of quantization accuracy information are encoded.
  • the N pieces of quantization accuracy information, which are encoded using the selected quantization accuracy information encoding method, and each of the band signals are multiplexed. Accordingly, the encoding efficiency can be improved.
  • the multiplexing of at least one band signal with quantization accuracy information and with an encoded normalization coefficient for each band signal is demultiplexed.
  • the demultiplexed normalization coefficient is decoded.
  • Each band signal is dequantized on the basis of the quantization accuracy information, and the dequantized signal is denormalized on the basis of the normalization coefficient, whereby an acoustic time-series signal is generated. Bands of generated band signals are combined. Accordingly, the encoding efficiency can be improved.
  • the multiplexing of at least one band signal with a normalization coefficient and with encoded quantization accuracy information for each band signal is demultiplexed.
  • the demultiplexed quantization accuracy information is decoded.
  • Each band signal is dequantized on the basis of the quantization accuracy information, and the dequantized signal is denormalized on the basis of the normalization coefficient, whereby an acoustic time-series signal is generated. Bands of generated band signals are combined. Accordingly, the encoding efficiency can be improved.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Signal Processing (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • Computational Linguistics (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Human Computer Interaction (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Compression, Expansion, Code Conversion, And Decoders (AREA)
US10/204,298 2000-12-22 2001-12-21 Encoder and decoder Expired - Lifetime US7016502B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US11/324,633 US7933417B2 (en) 2000-12-22 2006-01-03 Encoding device and decoding device

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
JP2000-390589 2000-12-22
JP2000390589 2000-12-22
PCT/JP2001/011254 WO2002052732A1 (en) 2000-12-22 2001-12-21 Encoder and decoder

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US11/324,633 Continuation US7933417B2 (en) 2000-12-22 2006-01-03 Encoding device and decoding device

Publications (2)

Publication Number Publication Date
US20030112979A1 US20030112979A1 (en) 2003-06-19
US7016502B2 true US7016502B2 (en) 2006-03-21

Family

ID=18856922

Family Applications (2)

Application Number Title Priority Date Filing Date
US10/204,298 Expired - Lifetime US7016502B2 (en) 2000-12-22 2001-12-21 Encoder and decoder
US11/324,633 Expired - Fee Related US7933417B2 (en) 2000-12-22 2006-01-03 Encoding device and decoding device

Family Applications After (1)

Application Number Title Priority Date Filing Date
US11/324,633 Expired - Fee Related US7933417B2 (en) 2000-12-22 2006-01-03 Encoding device and decoding device

Country Status (6)

Country Link
US (2) US7016502B2 (ko)
EP (1) EP1345331B1 (ko)
JP (2) JP4300800B2 (ko)
KR (1) KR100844810B1 (ko)
DE (1) DE60135487D1 (ko)
WO (1) WO2002052732A1 (ko)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040158456A1 (en) * 2003-01-23 2004-08-12 Vinod Prakash System, method, and apparatus for fast quantization in perceptual audio coders
US11289102B2 (en) * 2013-12-02 2022-03-29 Huawei Technologies Co., Ltd. Encoding method and apparatus

Families Citing this family (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1345331B1 (en) * 2000-12-22 2008-08-20 Sony Corporation Encoder
JP4296752B2 (ja) 2002-05-07 2009-07-15 ソニー株式会社 符号化方法及び装置、復号方法及び装置、並びにプログラム
EP1989707A2 (fr) * 2006-02-24 2008-11-12 France Telecom Procede de codage binaire d'indices de quantification d'une enveloppe d'un signal, procede de decodage d'une enveloppe d'un signal et modules de codage et decodage correspondants
JP5098453B2 (ja) * 2007-06-12 2012-12-12 カシオ計算機株式会社 音声符号化装置、音声復号装置、音声符号化方法、音声復号方法、及び、プログラム
JP5098458B2 (ja) * 2007-06-20 2012-12-12 カシオ計算機株式会社 音声符号化装置、音声符号化方法、及び、プログラム
JP5434592B2 (ja) * 2007-06-27 2014-03-05 日本電気株式会社 オーディオ符号化方法、オーディオ復号方法、オーディオ符号化装置、オーディオ復号装置、プログラム、およびオーディオ符号化・復号システム
CA2858944C (en) * 2007-11-12 2017-08-22 The Nielsen Company (Us), Llc Methods and apparatus to perform audio watermarking and watermark detection and extraction
US8457951B2 (en) 2008-01-29 2013-06-04 The Nielsen Company (Us), Llc Methods and apparatus for performing variable black length watermarking of media
US8700410B2 (en) * 2009-06-18 2014-04-15 Texas Instruments Incorporated Method and system for lossless value-location encoding
US8755432B2 (en) 2010-06-30 2014-06-17 Warner Bros. Entertainment Inc. Method and apparatus for generating 3D audio positioning using dynamically optimized audio 3D space perception cues
US8917774B2 (en) * 2010-06-30 2014-12-23 Warner Bros. Entertainment Inc. Method and apparatus for generating encoded content using dynamically optimized conversion
US9591374B2 (en) 2010-06-30 2017-03-07 Warner Bros. Entertainment Inc. Method and apparatus for generating encoded content using dynamically optimized conversion for 3D movies
US10326978B2 (en) 2010-06-30 2019-06-18 Warner Bros. Entertainment Inc. Method and apparatus for generating virtual or augmented reality presentations with 3D audio positioning
ES2765527T3 (es) 2011-04-20 2020-06-09 Panasonic Ip Corp America Dispositivo y método para la ejecución de la codificación de Huffman
CN106941004B (zh) * 2012-07-13 2021-05-18 华为技术有限公司 音频信号的比特分配的方法和装置
EP3961622B1 (en) 2013-05-24 2023-11-01 Dolby International AB Audio encoder
EP2980793A1 (en) * 2014-07-28 2016-02-03 Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. Encoder, decoder, system and methods for encoding and decoding

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1990009022A1 (en) 1989-01-27 1990-08-09 Dolby Laboratories Licensing Corporation Low bit rate transform coder, decoder and encoder/decoder for high-quality audio
JPH0854899A (ja) 1994-08-11 1996-02-27 Sharp Corp 符号化装置
JPH08181617A (ja) 1994-12-21 1996-07-12 Sony Corp ディジタル信号伝送装置
US5731767A (en) * 1994-02-04 1998-03-24 Sony Corporation Information encoding method and apparatus, information decoding method and apparatus, information recording medium, and information transmission method
US5901234A (en) * 1995-02-14 1999-05-04 Sony Corporation Gain control method and gain control apparatus for digital audio signals
US5974379A (en) * 1995-02-27 1999-10-26 Sony Corporation Methods and apparatus for gain controlling waveform elements ahead of an attack portion and waveform elements of a release portion

Family Cites Families (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
IT1184023B (it) 1985-12-17 1987-10-22 Cselt Centro Studi Lab Telecom Procedimento e dispositivo per la codifica e decodifica del segnale vocale mediante analisi a sottobande e quantizzazione vettorariale con allocazione dinamica dei bit di codifica
JPH02153639A (ja) 1988-12-06 1990-06-13 Fujitsu Ltd 帯域分割正規化制御方式
JPH02288739A (ja) 1989-04-28 1990-11-28 Fujitsu Ltd 音声符号復号化伝送方式
JP2560873B2 (ja) 1990-02-28 1996-12-04 日本ビクター株式会社 直交変換符号化復号化方法
JP3186290B2 (ja) * 1993-01-20 2001-07-11 ソニー株式会社 符号化方法、符号化装置、復号化装置及び記録媒体
JP3277677B2 (ja) * 1994-04-01 2002-04-22 ソニー株式会社 信号符号化方法及び装置、信号記録媒体、信号伝送方法、並びに信号復号化方法及び装置
US6167093A (en) * 1994-08-16 2000-12-26 Sony Corporation Method and apparatus for encoding the information, method and apparatus for decoding the information and method for information transmission
JP3371590B2 (ja) 1994-12-28 2003-01-27 ソニー株式会社 高能率符号化方法及び高能率復号化方法
US5706009A (en) * 1994-12-29 1998-01-06 Sony Corporation Quantizing apparatus and quantizing method
JPH08263099A (ja) 1995-03-23 1996-10-11 Toshiba Corp 符号化装置
JPH09135173A (ja) 1995-11-10 1997-05-20 Sony Corp 符号化装置および符号化方法、復号化装置および復号化方法、伝送装置および伝送方法、並びに記録媒体
JP3344944B2 (ja) 1997-05-15 2002-11-18 松下電器産業株式会社 オーディオ信号符号化装置，オーディオ信号復号化装置，オーディオ信号符号化方法，及びオーディオ信号復号化方法
JP3246715B2 (ja) 1996-07-01 2002-01-15 松下電器産業株式会社 オーディオ信号圧縮方法，およびオーディオ信号圧縮装置
JP3998281B2 (ja) 1996-07-30 2007-10-24 株式会社エイビット デジタル音声信号の帯域分割符号化方法と復号化方法
US5924064A (en) * 1996-10-07 1999-07-13 Picturetel Corporation Variable length coding using a plurality of region bit allocation patterns
JP2000267699A (ja) 1999-03-19 2000-09-29 Nippon Telegr & Teleph Corp <Ntt> 音響信号符号化方法および装置、そのプログラム記録媒体、および音響信号復号装置
JP2001094433A (ja) * 1999-09-17 2001-04-06 Matsushita Electric Ind Co Ltd サブバンド符号化・復号方法
EP1345331B1 (en) * 2000-12-22 2008-08-20 Sony Corporation Encoder

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1990009022A1 (en) 1989-01-27 1990-08-09 Dolby Laboratories Licensing Corporation Low bit rate transform coder, decoder and encoder/decoder for high-quality audio
JPH04504192A (ja) 1989-01-27 1992-07-23 ドルビー・ラボラトリーズ・ランセンシング・コーポレーション 高品質オーディオ用低ビットレート変換コーダ、デコーダ、及びエンコーダ・デコーダ
US5731767A (en) * 1994-02-04 1998-03-24 Sony Corporation Information encoding method and apparatus, information decoding method and apparatus, information recording medium, and information transmission method
JPH0854899A (ja) 1994-08-11 1996-02-27 Sharp Corp 符号化装置
JPH08181617A (ja) 1994-12-21 1996-07-12 Sony Corp ディジタル信号伝送装置
US5901234A (en) * 1995-02-14 1999-05-04 Sony Corporation Gain control method and gain control apparatus for digital audio signals
US5974379A (en) * 1995-02-27 1999-10-26 Sony Corporation Methods and apparatus for gain controlling waveform elements ahead of an attack portion and waveform elements of a release portion

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040158456A1 (en) * 2003-01-23 2004-08-12 Vinod Prakash System, method, and apparatus for fast quantization in perceptual audio coders
US7650277B2 (en) * 2003-01-23 2010-01-19 Ittiam Systems (P) Ltd. System, method, and apparatus for fast quantization in perceptual audio coders
US11289102B2 (en) * 2013-12-02 2022-03-29 Huawei Technologies Co., Ltd. Encoding method and apparatus

Also Published As

Publication number Publication date
US7933417B2 (en) 2011-04-26
US20030112979A1 (en) 2003-06-19
KR20030009369A (ko) 2003-01-29
EP1345331B1 (en) 2008-08-20
JP4300800B2 (ja) 2009-07-22
KR100844810B1 (ko) 2008-07-09
EP1345331A1 (en) 2003-09-17
JP2009058976A (ja) 2009-03-19
JP4844620B2 (ja) 2011-12-28
DE60135487D1 (de) 2008-10-02
WO2002052732A1 (en) 2002-07-04
JPWO2002052732A1 (ja) 2004-04-30
EP1345331A4 (en) 2007-05-02
US20060115092A1 (en) 2006-06-01

Similar Documents

Publication Publication Date Title
US7933417B2 (en) Encoding device and decoding device
EP1734511B1 (en) Entropy coding by adapting coding between level and run-length/level modes
US7627480B2 (en) Support of a multichannel audio extension
US6542863B1 (en) Fast codebook search method for MPEG audio encoding
US7848931B2 (en) Audio encoder
US20020049586A1 (en) Audio encoder, audio decoder, and broadcasting system
WO1995032499A1 (fr) Procede de codage, procede de decodage, procede de codage-decodage, codeur, decodeur et codeur-decodeur
JP2013250563A (ja) レベル・モードとラン・レングス／レベル・モードの間での符号化を適応させるエントロピー符号化
WO1995034956A1 (fr) Procede et dispositif de codage de signal, procede et dispositif de decodage de signal, support d&#39;enregistrement et dispositif de transmission de signaux
WO2007011157A1 (en) Virtual source location information based channel level difference quantization and dequantization method
US6601032B1 (en) Fast code length search method for MPEG audio encoding
JP3318931B2 (ja) 信号符号化装置、信号復号化装置及び信号符号化方法
US20020173969A1 (en) Method for decompressing a compressed audio signal
JPH0969781A (ja) オーディオデータ符号化装置
US7181079B2 (en) Time signal analysis and derivation of scale factors
US6678648B1 (en) Fast loop iteration and bitstream formatting method for MPEG audio encoding
JP2820096B2 (ja) 符号化方式および復号方式
JPH09135173A (ja) 符号化装置および符号化方法、復号化装置および復号化方法、伝送装置および伝送方法、並びに記録媒体
JP3998281B2 (ja) デジタル音声信号の帯域分割符号化方法と復号化方法
JP3389849B2 (ja) 量子化装置
JPH0645943A (ja) 音声符号化／復号化方式

Legal Events

Date Code Title Description
AS Assignment

Owner name: SONY CORPORATION, JAPAN

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:TOYAMA, KEISUKE;SUZUKI, SHIRO;TSUJI, MINORU;REEL/FRAME:013458/0048

Effective date: 20020830

STCF Information on status: patent grant

Free format text: PATENTED CASE

FEPP Fee payment procedure

Free format text: PAYOR NUMBER ASSIGNED (ORIGINAL EVENT CODE: ASPN); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY

FPAY Fee payment

Year of fee payment: 4

FPAY Fee payment

Year of fee payment: 8

MAFP Maintenance fee payment

Free format text: PAYMENT OF MAINTENANCE FEE, 12TH YEAR, LARGE ENTITY (ORIGINAL EVENT CODE: M1553)

Year of fee payment: 12